發生了什麼事
The BBC reports that Anthropic researcher Jacob Coxon resigned and publicly warned that advanced AI systems could become dangerous. His comments drew support from some Anthropic employees but also criticism from Silicon Valley executives and investors, including Grindr CEO George Arison, who said he told some engineers to stop using Anthropic’s technology.
The BBC reports that 27-year-old Jacob Coxon, who previously worked at OpenAI, abruptly resigned from Anthropic and wrote that people developing AI were “gambling with our lives.” The report also cites Anthropic team lead Evan Hubinger publicly estimating a greater-than-10% chance that AI could kill all humans within the next decade. These statements and the circumstances of Coxon’s departure were not independently confirmed in the source. Anthropic was asked for comment but had not responded when the report was published.
According to the BBC, the comments became a subject of discussion at a Goldman Sachs conference in San Francisco. Grindr chief executive George Arison called the warnings dangerous and said they led him to instruct some engineers to stop using Anthropic’s technology. The BBC also reports that Nvidia chief executive Jensen Huang dismissed the warnings in remarks described by multiple conference attendees, while investors Brad Gerstner and Hugging Face chief executive Clement Delangue criticized the rhetoric before later taking more measured positions.
為什麼這很重要
The dispute shows how disagreements over catastrophic AI risks are affecting business decisions, investor sentiment and policy debates. It also highlights the unresolved tension between warning about powerful systems and promoting their commercial value. The BBC’s report does not independently establish the probability of extinction, the motives behind any warning, or the broader effect on Anthropic’s customers.
The immediate significance is practical rather than merely rhetorical: a named technology user told the BBC that public statements about existential risk influenced an internal decision about using Anthropic’s systems. If similar reactions spread, safety messaging could affect procurement, partnerships and investor confidence, even without a change to model capabilities or availability.
The report also places the argument within a larger policy and financing dispute. The BBC says Anthropic and OpenAI are preparing for potentially major public offerings, while US lawmakers are considering a temporary pause on advanced AI development. Those financing and legislative developments, their valuations, and claims that safety warnings are intended to strengthen market positions were not independently established by the source.
互動機制：它實際上是如何運作的
以互動方式探索這項發展背後的基礎技術。
Why can ethical evaluation not be reduced to one model score?
接下來看什麼
The key developments are responses from Anthropic and OpenAI, any verified changes in enterprise use of Anthropic systems, and whether proposed US legislation advances. The report provides no documented product access or pricing change.
Watch for a direct Anthropic response to Coxon’s resignation and the warnings cited by the BBC, as well as any evidence that customers are changing procurement or deployment decisions. The source does not quantify how many organizations have reduced or stopped using Anthropic systems.
Watch whether the proposed Ban Artificial Superintelligence Act advances, changes substantially, or remains only a legislative proposal. The report also points to future securities filings and questions about profitability, but it does not establish filing dates, financial results, or whether either company will complete an initial public offering.